Generative AI Activity Monitoring

Investigate AI Use and Sensitive-Data Exposure

See who or what initiated an AI interaction, what data was sent, and where sensitive information appeared. Search activity, investigate events, and connect findings to alerts and protection policies.

Book a Demo Explore Integrations

OpenAI, Claude, Bedrock, Azure AI, and Gemini
Identity, Prompt, and Response Context
Search, Alerts, and Reports

Why teams monitor Generative AI

Know what enters and leaves AI services

AI services can receive credentials, personal data, financial records, health information, and other sensitive content. Security teams need a clear record of who or what sent it, what the service returned, and which policy matched.

DataSunrise records AI interactions with the provider, available user or application identity, prompts, responses, attachments when available, timing, and policy matches. Teams can search this activity, investigate sensitive-data exposure, and use the findings to improve alerts, masking, and prompt controls.

How AI monitoring works

From AI interaction to searchable evidence

DataSunrise captures the available identity and content context, checks it against sensitive-data policies, and records a trail teams can investigate.

Capture interaction context

Available user or application identity
AI service and model
Prompt, response, and available attachments

Check for sensitive data

Information Type matching
Policy matches
Timing and event outcome

Investigate and respond

Searchable activity records
Alerts and reports
Links to masking and prompt controls

Security teams can trace an AI interaction from the sender and content to the policy match, investigation, and next action.

Product evidence

See the workflow in DataSunrise

Product screens show how policy, evidence, and protection appear in practice.

DataSunrise Add Database screen with Generative AI selected as the connection type and fields for the logical name, endpoint, port, authentication, and certificate settings — **Connect a Generative AI service** DataSunrise configures a Generative AI service as a monitored connection. Teams define the endpoint, port, authentication, and connection settings before applying audit or protection policies.

DataSunrise Audit Rule Details screen with Generative AI selected as the database type and a Generative AI instance added to the rule — **Apply an audit rule to AI activity** An Audit rule selects the Generative AI connection whose interactions should be recorded. The rule can then add filters, evidence settings, alerts, and reporting behavior.

DataSunrise Transactional Trails showing a Generative AI session and a recorded prompt event with its rule, instance, time, row count, error status, and query type — **Review recorded AI interactions** Transactional Trails place Generative AI activity in a searchable event list. This example records the session and prompt event with the matching rule, target instance, time, and outcome.

1 / 3

Generative AI monitoring capabilities

Explore Generative AI Activity Monitoring

1 / 4

See AI Activity in One Place

Bring activity from OpenAI and ChatGPT, Anthropic Claude, Amazon Bedrock, Azure OpenAI and Azure AI, and Google Gemini into one investigation workflow.

Security teams can review AI use without piecing together separate provider logs. The integration directory shows the connection options for each service.

Find Sensitive Data in AI Conversations

DataSunrise checks prompts, responses, and available attachments for the sensitive information your organization defines, including personal, financial, health, and credential data.

Policy matches show investigators where protected information appeared and which interactions need attention.

Investigate Who Sent What

Search AI activity by provider, available user or application identity, time, and policy match. Review the prompt, response, and attachment context available for the event.

Alerts and reports help teams investigate incidents, review AI use, and keep evidence for security or compliance work.

Connect Visibility to AI Protection

Monitoring shows what happened and creates evidence for investigation. Generative AI Data Masking and Prompt Controls use configured policies to mask sensitive data or block risky prompts.

Teams can use monitoring results to refine those controls and focus them on the interactions that carry the most risk.

Common scenarios

Where Teams Use It

Trace sensitive prompts
Find interactions where credentials, personal data, or other protected content appeared.
Investigate user and application activity
See who or what initiated an interaction and review the available prompt and response context.
Support audit and compliance review
Keep searchable activity, policy matches, alerts, and reports for later review.
Improve AI protection policies
Use monitoring evidence to refine masking and prompt controls.

FAQ

Frequently Asked Questions

Which AI services can DataSunrise monitor?

DataSunrise monitors OpenAI and ChatGPT, Anthropic Claude, Amazon Bedrock, Azure OpenAI and Azure AI, and Google Gemini.

How does DataSunrise connect to AI services?

DataSunrise uses Squid and eCAP to connect AI traffic to its monitoring, masking, and prompt-control policies.

What is the difference between AI monitoring and AI protection?

Monitoring records activity and alerts teams to policy matches. Generative AI Data Masking and Prompt Controls can mask sensitive data or block risky prompts.

Map the right workflow

See how DataSunrise works with your technology stack

View Integration Examples